INDEX

Explanations

phrases that depict learning from others and building upon shared knowledge

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

BuyableInstoreAndOnline

-0.75

portation

-0.68

stead

-0.67

pool

-0.64

requ

-0.61

................................

-0.61

 subsidy

-0.61

 redemption

-0.60

 cancell

-0.60

................................................................

-0.60

POSITIVE LOGITS

»Ĵ

0.96

 glance

0.92

 analys

0.87

 analyzing

0.85

 hindsight

0.84

 studying

0.81

FAC

0.79

 hier

0.78

 subconscious

0.78

 decipher

0.77

Activations Density 0.485%

phrases that depict learning from others and building upon shared knowledge

No Comments

No Known Activations